Fast Follower
A fast follower is a company that doesn't bring a new idea to market first, but shortly afterward launches its own, often improved version. In the AI industry, this is one of the most important strategies, because copying is much cheaper than inventing.
Whoever brings a new product to market first has a head start. But they also bear all the costs and all the failures alone. A fast follower deliberately chooses second place: it waits until someone else shows that an idea works, and then quickly brings out its own version. The goal is not to slowly catch up, but to close the gap within months. Often the second version is cheaper, easier to use, or built into products that already exist. The term comes from economics and is nowadays especially often used with regard to technology companies.
The Advantage of Being Second
The first mover has to guess what customers want. It develops in a direction that may turn out to be a dead end. These failed attempts cost money that nobody gets back. A fast follower, on the other hand, can already see which variant has prevailed. Experts also call this advantage the second-mover advantage.
On top of that comes a difference in cost. Basic research is expensive; rebuilding a known solution is significantly cheaper. Studies estimate the effort for imitators, depending on the industry, at a fraction of the original development costs. In the case of medicines, patents often protect the inventor for twenty years. For software and AI methods, this kind of protection usually doesn’t exist, because ideas are published in academic papers.
The disadvantage is obvious: whoever only ever follows never sets the direction. And sometimes the first mover has grown so quickly that users no longer want to switch at all. This effect is called lock-in, meaning being stuck with one provider.
How Catching Up Works in Practice
The first step is observation. Large companies read research papers, test competitors' products, and analyze what users criticize about them. Especially in AI research, a great deal is published, sometimes even including the source code. Whoever knows the method doesn’t need to reinvent it.
The second step is rebuilding it with one’s own resources. A fast follower uses what it already has: data centers, data stores, development teams, and millions of existing customers. One well-known technique is distillation. In this process, a smaller model is taught to imitate the answers of a larger role model. The result is cheaper to run and, in many tasks, nearly as good.
The third step is distribution. Building a new product into an app that is already widely used immediately provides millions of users. This is exactly where the inventor often loses its lead. It has the better lab, while the fast follower has the bigger distribution network.
Fast Followers in the AI Industry
The best-known example is chatbots. OpenAI released ChatGPT in late 2022 and was clearly the first mover. Within about a year, Google, Meta, Anthropic, and several Chinese companies had comparable systems. None of these followers had to discover the basic idea themselves.
The same pattern played out with image generators and voice assistants. One provider shows that a feature delights users. A few months later, it’s built into search engines, office software, and smartphones. The term comes up in stock market news when analysts ask whether a company is losing its lead or is simply saving cleverly.
A common mistake is to equate fast followers with copycats. Pure copying without any improvement of one’s own rarely works, because customers need a reason to switch. Successful fast followers make the product faster, cheaper, or combine it with something only they offer. Well-known examples outside AI are Samsung after the first iPhone, or Facebook, which adopted Snapchat’s short videos.